[cp-patches] Re: Some VMStackWalker cleanups/simplifications

[Robert Lougher]

Hi List,

Archie Cobbs <archie <at> dellroad.org> writes:

> I'd like to better understand the issue with those VMs. It seems to
> me that VMStackBrowser and VMStackWalker can both be implemented in
> terms of the other (one's an iterator, one returns an array). So how
> can a VM be able to implement one API but not the other?
> 

I think you're getting a little confused with the VMStackBrowser patch that was 
posted around the same time.  I believe this was proposed without any reference 
to that.  The comment at the bottom about VMStackBrowser is a "get out" because 
the other patch was posted too soon to be considered.
 
> Or is this just a matter of convenience, as one way is closer to
> the way those particular VM's already do it? If that's the case then
> we're getting into the murky waters of VM-specific "convenience code".
> 
> I agree passing the Class parameter is a bit ugly in any case.
> 

See my other post for why I think the Class parameter was proposed.  
Personally, I think the array copy in SecurityManager.getClassContext() is even 
uglier.

Rob.